Upload 6 files

Dockerfile

@@ -25,14 +25,14 @@ COPY . .
 # Setup environment variables for compilation
 # We need to find where pip installed llama-cpp-python to link against it
 # In docker, it's usually /usr/local/lib/python3.12/site-packages
-# ENV SITE_PACKAGES=/usr/local/lib/python3.12/site-packages
+ENV SITE_PACKAGES=/usr/local/lib/python3.12/site-packages
 
 # Compile the engine
 # WORKDIR /app/engine
-# RUN g++ -O2 -shared -fPIC -o libbatch.so batch_server.cpp \
-#     -I"${SITE_PACKAGES}/include" \
-#     -L"${SITE_PACKAGES}/llama_cpp/lib" \
-#     -lllama -Wl,-rpath,"${SITE_PACKAGES}/llama_cpp/lib"
+RUN g++ -O2 -shared -fPIC -o libbatch.so batch_server.cpp \
+    -I"${SITE_PACKAGES}/include" \
+    -L"${SITE_PACKAGES}/llama_cpp/lib" \
+    -lllama -Wl,-rpath,"${SITE_PACKAGES}/llama_cpp/lib"
 
 # Setup Model (Download during build or mount volume? 
 # Best practice: Download in build if small, or use script at runtime.